The Presented topics were:
Schwarz, J. A. (2018): Analyzing the impact of finite capacity on sales and operations planning for product rollovers. EURO 2018, Valencia, Spain, July 2018.
A product rollover takes place if a product generation is replaced by a successor version. This occurs frequently in industries which are characterized by short product life cycles such as semiconductor, consumer electronics, and fashion. We focus on the questions whether and how the presence of limited production capacities drives the selection of rollover strategies and the underlying optimal decisions on production, sales, and price. We propose a deterministic two-period model
of a profit-maximizing company. A vertical demand model is used to capture assortment based substitution and stock-out based substitution.
The problem is formalized as a non-linear integer program. We provide explicit solutions for the unlimited capacity case and closed-form solutions for given rollover strategies for the finite capacity case with customers that are unwilling to substitute in case of stock outs. Numerical results are obtained for the general setting. We show that the selected limited production capacity drives the firm to offer the old and new product generations simultaneously. Moreover, a decreasing willingness to substitute in case of stock-outs can cause the postponement of the introduction of the new products. Finally limited productioncapacity can lead to increased discounts for the older product, i.e., a lower price.
Schnitzler, J. M. and R. Stolletz (2018): A branch and bound procedure for the stochastic assembly line balancing problem. EURO 2018, Valencia, Spain, July 2018.
We analyze the assembly line balancing problem where tasks have to be assigned to stations with the goal of minimizing the number of stations used. Task times are stochastic, leading to the possibility of incomplete work pieces. Therefore, we consider a constraint on the probability of finishing a work piece within the given cycle time.
A sampling model is developed to account for generally distributed task times. We present a bidirectional branch and bound procedure in order to solve the model to optimality. A numerical study compares the performance of the algorithm to the solution with standard solvers.